Long-Term Myoelectric Training with Delayed Feedback in the Home Environment

Area of Science:

• Rehabilitation Engineering
• Neuroscience
• Human-Computer Interaction

Background:

• Myoelectric prosthesis users often lack immediate sensory feedback from their devices.
• Effective control requires users to generate distinct muscle activations without real-time input.
• Previous studies using real-time feedback leave ambiguity regarding short-term adaptation versus motor learning.

Purpose of the Study:

• To investigate if motor learning, rather than short-term adaptation, drives performance improvements in myoelectric control.
• To determine if delayed feedback can facilitate motor learning for prosthesis control.
• To assess the efficacy of abstract decoding with delayed feedback for enabling multi-class control.

Main Methods:

• A delayed feedback paradigm was employed, with feedback presented after trial completion (~1.5 s).
• Three participants trained over five days in their home environments.
• A total of 4920 trials were completed across all participants.

Main Results:

• Participants achieved high accuracy despite receiving no real-time control input feedback.
• Performance gains were retained across multiple training days.
• The delayed feedback paradigm successfully facilitated learning.

Conclusions:

• Delayed feedback in abstract decoding promotes motor learning in myoelectric prosthesis control.
• This approach enables robust four-class control without the need for immediate sensory feedback.
• Findings suggest a viable training strategy for improving prosthesis usability and user adaptation.